On May 13, 2026, South Korea's Korea Communications Commission (KCC) published the enforcement decree that will give teeth to the amended Act on Promotion of Information and Communications Network Utilization — the statute critics have nicknamed the "False Information Eradication Act." The amendment cleared the National Assembly on December 24, 2025 and was approved by the State Council on December 30; both the law and its decree take effect in July 2026. Public comment on the decree ran through May 27. The package sets new content-moderation duties for large platforms — social networks, video services, search, and marketplaces above a daily-user threshold set by presidential decree (reported at roughly one million average daily users).
The case Seoul is making
The strongest version of the government's argument deserves a fair hearing. Generative AI has made convincing deepfakes cheap, election seasons have been polluted by fabricated audio and video, and victims of manipulated content often have no fast remedy. Notably, the law's headline takedown duty is comparatively narrow: it compels platforms to remove content and suspend accounts only once a court has confirmed the material is illegal or false. Tying removal to a judicial finding is more defensible than the notice-and-takedown-on-demand regimes seen elsewhere, and mandatory twice-yearly transparency reports — disclosing user numbers, report volumes, and actions taken — are a genuine accountability gain. On the merits of court-confirmed content, this is a real harm meeting a real process.
Where proportionality breaks down
The trouble starts beyond the court-order core. The amendment also requires large providers to adopt and run their own "self-regulation" policies against false and manipulated information, backed by a penalty surcharge of up to 1 billion won (about $684,000) for repeatedly distributing content a court has deemed false. As the Korean outlet Newsspace and legal analysts have flagged, the operative definition — content "wholly or partially untrue" distributed with intent and purpose to harm — is broad and contestable. "Partially untrue" is doing enormous work in a sentence that can reach satire-adjacent commentary, contested scientific claims, and ordinary political argument.
Vague standards plus financial penalties produce a predictable incentive: when in doubt, take it down. A platform facing a six-figure surcharge and reputational exposure will not litigate every borderline post; it will over-remove. That is the collateral-censorship dynamic documented across content-liability regimes worldwide — the cost of an erroneous removal falls on a silent user, while the cost of leaving content up falls on the company's balance sheet. The asymmetry, not the intent, is what chills speech.
The Transparency Center problem
The decree also stands up a new Information and Communications Service Transparency Center to set standards for and support fact-checking organizations. The instinct — building institutional capacity to assess claims — is reasonable. But a fact-checking apparatus convened and underwritten by the communications regulator risks becoming, in critics' words, a "government-friendly" arbiter of truth. Independent fact-checking earns trust precisely because it is independent of the state. Once a regulator funds the checkers and certifies their standards, the line between neutral verification and official viewpoint enforcement blurs — and so does the credibility of every verdict the Center touches.
The cross-border dimension
Because the rules bind any large platform serving Korean users, they land squarely on American firms. The U.S. State Department registered "substantial concerns" on December 31, 2025, warning the law harms U.S. platform businesses and "weakens freedom of expression"; Deputy Assistant Secretary Sarah Rogers argued that deepfakes are better addressed through civil remedies to victims than through "invasive license for viewpoint-based censorship." The trade track is active too: USTR's 2026 National Trade Estimate, released March 31, 2026, catalogued a range of Korean digital measures, and in April Trade Representative Jamieson Greer told Congress Washington wants "outcomes" from Korea and others on regulations affecting U.S. platforms such as Google, Meta, and X. Whatever one thinks of Section 301 as a tool, the friction is a signal that Korea's drafting is broad enough to read as a trade and speech problem, not just a domestic safety measure.
A more proportionate path
None of this requires Korea to abandon the effort. The court-confirmed takedown duty and transparency reporting are the parts worth keeping. The fixes are narrowing, not nullifying. First, tighten the definition: limit liability to demonstrably false statements of fact made with actual malice, and explicitly carve out opinion, satire, and good-faith reporting beyond the existing parody exception. Second, replace open-ended "self-regulation" surcharges with a duty to maintain due-process machinery — clear notice, a genuine appeal, and reinstatement when a removal proves wrong — rather than penalties that reward deletion. Third, sever the Transparency Center from the regulator's editorial reach, funding fact-checking through an arm's-length body with published methodology and no power to compel outcomes.
Disinformation is a real harm, and a democracy is entitled to respond. But the measure of a good rule is whether it removes the lie without deterring the inconvenient truth. As written, Korea's decree gets the hard, court-anchored part right while leaving the easy levers — vague terms, removal-biased penalties, and a state-linked truth office — pointed in the wrong direction. Seoul has six months and a comment record to fix that before July.